🔢 mathematics

On higher Du Bois singularities and $K$ -regularity

This paper investigates the relationship between higher Du Bois singularities and $K$ -regularity, utilizing vanishing theorems for Du Bois complexes to prove a strengthened version of Vorst's conjecture for local complete intersections in characteristic zero and to construct new examples illustrating phenomena in $K$ -regularity.

Original authors: Wanchun Shen

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Wanchun Shen

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are an architect inspecting a building. If the building is perfectly constructed with straight lines and smooth corners, we call it "regular." But what if the building has cracks, jagged edges, or weird, twisted corners? In the world of algebraic geometry, these "buildings" are mathematical shapes called varieties, and their "cracks" are called singularities.

This paper, written by Wanchun Shen, is about a specific way of measuring how "broken" or "irregular" these shapes are. The author uses two different tools to measure the damage: one comes from a field called Algebraic K-theory (let's call this the "K-regularity" tool), and the other comes from Hodge theory (a branch of math dealing with shapes and waves, which the author calls "Du Bois singularities").

Here is the breakdown of the paper's main ideas, explained simply:

1. The Two Measuring Rulers

The paper tries to connect two different ways of checking if a shape is "good" (regular) or "bad" (singular).

Ruler A: K-regularity. Imagine you have a shape, and you try to stretch it out infinitely in a new direction (like adding a long hallway to a room). If the shape behaves perfectly and doesn't change its "identity" when you do this, it is considered K-regular. If it breaks or changes weirdly, it's not. The paper asks: How broken does a shape have to be before it stops being K-regular?
Ruler B: Du Bois Singularities. This is a more technical way of looking at the "texture" of the cracks. Mathematicians have developed a scale called m-Du Bois. If a shape is "m-Du Bois," it means its cracks are "smooth enough" up to a certain level of complexity ( $m$ ). Think of it like grading a scratch on a car: a light scratch is "low m," while a deep, jagged tear is "high m."

2. The Big Discovery: Connecting the Rulers

The main achievement of this paper is building a bridge between Ruler A and Ruler B. The author proves that you can predict how "K-regular" a shape is just by looking at its "Du Bois" texture.

The Analogy: It's like realizing that if you know exactly how deep the cracks are in a bridge (Du Bois), you can instantly calculate whether the bridge will hold up under a specific type of stress test (K-regularity).
The Result: The paper provides a precise formula. For example, if a shape has a certain "minimal exponent" (a number that measures the severity of the crack), the author can tell you exactly which level of K-regularity the shape possesses.

3. Solving an Old Puzzle (Vorst's Conjecture)

There was a long-standing guess in the math world called Vorst's Conjecture. It asked: If a shape is K-regular up to a certain high level, does that mean the shape is actually perfect (smooth) with no cracks at all?

Previous Knowledge: Mathematicians knew this was true for some simple shapes.
This Paper's Contribution: The author proves a stronger version of this conjecture. They show that for a specific class of shapes (called "local complete intersections," which are shapes built by intersecting smooth surfaces), if the shape passes a relatively low-level K-regularity test (specifically, being K2-regular), then the shape must be perfectly smooth.
The Metaphor: It's like saying, "If a car passes a very basic safety inspection (K2-regular), we can be 100% sure it has no hidden structural damage at all."

4. New Tools and Examples

The paper doesn't just prove theorems; it builds new tools to create examples.

The "Minimal Exponent": The author uses a number (the minimal exponent) to act as a "severity meter" for singularities. By calculating this number, they can generate examples of shapes that are "just barely" K-regular or "just barely" broken.
Projective Varieties: The paper also looks at shapes that are "closed" (like a sphere) rather than "open" (like a flat plane). It shows that for these closed shapes, K-regularity depends on how certain mathematical waves (cohomology) behave inside the shape.

5. The "Bass Question"

The paper also tackles a question posed by mathematician Hyman Bass: If a shape is stable in one direction (A1-invariance), is it automatically stable in two directions (A2-invariance)?

The author uses the new connection between K-regularity and Du Bois singularities to create examples where the answer is Yes and examples where the answer is No. This helps map out the boundaries of when this rule holds true.

Summary

In short, Wanchun Shen's paper is a "translation guide" between two different languages used to describe broken mathematical shapes. By translating the "texture of the cracks" (Du Bois) into "stress-test results" (K-regularity), the author:

Proves that for many shapes, passing a moderate stress test guarantees the shape is actually perfect.
Provides a formula to calculate exactly how "broken" a shape is based on its geometry.
Creates new examples to show where these rules work and where they break down.

The paper relies heavily on advanced "vanishing theorems" (mathematical rules that say certain complex numbers become zero under specific conditions) to make these connections work, but the core message is about finding a precise link between the geometry of a shape's cracks and its algebraic stability.

Technical Summary: On Higher Du Bois Singularities and K-Regularity

Problem Statement
The paper addresses the problem of characterizing $K$ -regularity for algebraic varieties, specifically the extent to which algebraic K-groups $K_m(X)$ fail to be $\mathbb{A}^1$ -invariant. A scheme $X$ is defined as $K_m$ -regular if the natural maps $K_m(X) \to K_m(X \times \mathbb{A}^r)$ are isomorphisms for all $r \geq 0$ . While regular schemes are $K_m$ -regular for all $m$ , the behavior of singular schemes is more complex. The central problem is to establish a precise relationship between the geometric nature of singularities (specifically higher Du Bois singularities) and the algebraic property of $K$ -regularity. This includes investigating Vorst's conjecture, which posits that for affine schemes of dimension $d$ , regularity is equivalent to $K_{d+1}$ -regularity, and seeking strengthened versions of this conjecture for specific classes of singularities.

Methodology
The author employs a synthesis of Hodge-theoretic methods and algebraic K-theory. The primary technical framework involves:

Du Bois Complexes and cdh-differentials: Utilizing the Du Bois complexes $\underline{\Omega}^p_{X/F}$ (derived from Deligne's mixed Hodge theory) and their relationship to cdh-differentials $\Omega^p_{cdh, X/F}$ . The paper relies on the isomorphism between these objects established by Cortiñas, Haesemeyer, Schlichting, Walker, and Weibel (CHSW).
Higher Du Bois Singularities: Extending the classical notion of Du Bois singularities to "higher" analogs. A local complete intersection (lci) $X$ has $m$ -Du Bois singularities if the natural maps $\Omega^p_{X/F} \to \underline{\Omega}^p_{X/F}$ are isomorphisms for $p \leq m$ . For general varieties, the paper uses the notion of "pre- $m$ -Du Bois" singularities, defined by the vanishing of cohomology sheaves $H^i\underline{\Omega}^p_{X/F}$ for $i > 0$ and $p \leq m$ .
Vanishing Theorems: The proofs rely heavily on vanishing theorems for the cohomology of Du Bois complexes (e.g., Steenbrink's vanishing, Mustat¸˘a-Popa's codimension bounds) and results regarding the torsion and reflexivity of Kähler differentials for lci schemes (Graf, Miller-Vassiliadou).
K-Regularity via NK-groups: The study of $K$ -regularity is reduced to the study of Bass' $NK$-groups via the local-to-global spectral sequence. The paper utilizes the decomposition of $NK$-groups in terms of the cohomology of the cones of the maps $L^p_{X/\mathbb{Q}} \to \Omega^p_{X/\mathbb{Q}}$ , where $L$ is the cotangent complex.

Key Contributions and Results

Relationship between Higher Du Bois and K-Regularity (Theorem C):
The paper establishes a bidirectional relationship between higher Du Bois singularities and $K$ -regularity.
- If a variety $X$ of dimension $d$ has Du Bois singularities, it is $K_{-d+1}$ -regular.
- If $X$ is pre- $m$ -Du Bois, it is $K_{-t}$ -regular for specific bounds on $t$ depending on the dimension of the singular locus.
- Conversely, for affine varieties (either lci or satisfying depth conditions on Kähler differentials), $K$ -regularity implies pre- $m$ -Du Bois singularities.
Numerical Characterization via Minimal Exponent (Theorem A):
For a complex hypersurface $X$ of dimension $d$ with isolated singularities, the paper provides a numerical criterion for $K_m$ -regularity using Saito's minimal exponent $\tilde{\alpha}_X$ :
$X \text{ is } K_m\text{-regular} \iff \tilde{\alpha}_X \geq \left\lceil \frac{m+d}{2} \right\rceil.$
This allows for the explicit detection of $K$ -regularity for specific singularities, such as the hypersurface $x^2 + y^3 + z^7 = 0$ .
Strengthened Vorst's Conjecture (Theorem B):
The paper proves that for an affine local complete intersection, $K_2$ -regularity implies regularity. This strengthens Vorst's original conjecture (which required $K_{d+1}$ -regularity) and previous extensions (which required $K_2$ -regularity only for normal varieties or specific codimension bounds). The proof utilizes the codimension bound on the singular locus for $m$ -Du Bois singularities ( $\text{codim } X_{\text{sing}} \geq 2m+1$ ) combined with the regularity criteria derived from the vanishing of NK-groups.
Projective Varieties and Local vs. Global Regularity (Theorem D):
The paper characterizes $K_m$ -regularity for projective varieties via the degeneration of the Hodge-to-de Rham spectral sequence. Specifically, $X$ is $K_m$ -regular if and only if the maps $H^i(X, L^p_{X/F}) \to H^i(X, \Omega^p_{X/F})$ are isomorphisms for $i-p \geq -m+1$ .
This result is used to construct examples (Example 7.4) showing that local $K_m$ -regularity does not imply global $K_m$ -regularity, highlighting a distinction not present in the affine case.
Bass' Question and Du Bois Invariants (Section 6):
The paper classifies Du Bois invariants ( $b_{p,q} = \dim H^q \underline{\Omega}^p$ ) for surfaces and uses these to address Bass' question: whether $K_m(X) \cong K_m(X \times \mathbb{A}^1)$ implies $K_m(X) \cong K_m(X \times \mathbb{A}^2)$ .
- For surfaces, the answer is positive if and only if $b_{0,1}$ and $b_{1,1}$ are both zero or both non-zero.
- The paper constructs examples of surfaces with arbitrary non-negative integer values for $b_{0,1}$ and $b_{1,1}$ , providing new instances where Bass' question has either a positive or negative answer.
- For homogeneous hypersurfaces with isolated singularities, the paper proves that Bass' question has a positive answer for non-positive K-groups.

Significance
The paper claims to provide a "strengthened form of Vorst's conjecture" for local complete intersections in characteristic zero, reducing the regularity threshold from $K_{d+1}$ to $K_2$ . By bridging Hodge theory (via Du Bois complexes and minimal exponents) with algebraic K-theory, the work offers new tools to construct examples of $K$ -regular varieties and to understand the failure of $\mathbb{A}^1$ -invariance. The results refine the understanding of how the geometry of singularities (specifically the codimension of the singular locus and the behavior of differential forms) dictates the algebraic K-theoretic properties of a scheme. The paper concludes by posing further questions regarding $K_2$ -regularity for non-affine varieties and the relationship between higher F-injectivity and K-regularity in positive characteristic.